AI Zone Admin Forum Add your forum

NEWS: Chatbots.org survey on 3000 US and UK consumers shows it is time for chatbot integration in customer service!read more..

Need advice for publishing AI articles
 
 

I’m really struggling to get some articles published, especially about my own AI architecture, especially since I don’t have access to typical software used in publications, so I could use some advice. Most of my problems stem from my low income, such as my inability to buy expensive software, inability to pay for travel expenses to present articles at conferences, inability to pay for publication fees in prestigious journals, and also from my not having an employer who will pay for such travel or fees or software. I haven’t published anything in years largely because of the effects of a bad economy, which forces me to keep changing jobs instead of doing anything of long-term benefit to anyone, namely publishing my own ideas. I’m using only the Windows operating system.

So here are some specific questions:

What free software would you recommend for generic circles/boxes/arrows type diagrams?
I’ve used Visio but I don’t own it and I didn’t like it anyway, I don’t have MS Word, and I have Paint but it’s ugly and awful and amateurish. I have MS Works but I suspect it doesn’t do diagrams like MS Word. I have only very limited access to MS Word on other people’s computers.

How safe and effective would it be to place my articles online, say in PDF format, especially for establishing documented precedence of my original ideas, without my having to publish them anywhere?
Mostly I’m concerned about people stealing my ideas. My whole idea of my publishing now is to document that I came up with those ideas first.

If such online articles are safe and effective, where would be the best place to such articles? In a blog? On my own web site? Attachments in AI forums? Are there other good options I don’t know about?

Are there any specific places you would particularly recommend for submission of articles on novel AGI/AI architectures?
Only recently I learned that there even exists an AGI conference and Singularity conferences, for example, so there could me many other places I simply haven’t heard of. I used to publish at neural network conferences regularly, but my architecture is starting to get so far away from neural networks that it might not even be applicable to that field anymore.

Which languages would you recommend for developing such novel architectures for demo purposes?
Again, the compilers/interpreters probably have to be free, or at least low cost. I’ve used Visual Python (= VPython) because it’s simple and has a simple graphical interface, though there are some drawbacks with it (it does only 3D, it won’t interface with inputted images, image clustering isn’t flexible enough, etc.). I have a free C++ compiler (Bloodshed) but it sounds like I will need to interface other free software with it to do graphics, and such interfacing is often problematic. Maybe Python would help with such an interface? My simulations will be extremely visually-oriented, though simple, like with moving colored dots, maybe clusters of colored dots, maybe simple geometrical figures like squares and lines.

Which software would you recommend for video capture, such as for placing demo simulation videos on YouTube?
I tried Debut, which is free, and it worked, but despite its recommendations I strongly suspect it had a virus. I got a warning message about virus-like behavior when I installed it, then I started getting viruses not long after installing it, then (so far) all such virus detections ceased after I uninstalled it. I might be willing to pay for such software to avoid such nonsense, but I couldn’t pay very much for it.

 

 
  [ # 1 ]

As daunting as the prospect might be, most of your problems would go away if you made the switch to Linux. Even if you are not financially constrained, making the switch to Linux will be hugely beneficial in the long run.

Even without Linux, I know of two software packages which will run on Windows and which may benefit you a great deal in the short term. One is Inkscape which is a 2D graphics and diagramming package based on SVG which is the graphics dialect of XML. It’s not just free, it’s brilliant!

The other package is from http://www.mendeley.com/ and it’s a reference manager and PDF organiser. Anyone who is doing serious research with a view to publishing their own work would find it indispensable.

As far as publishing your own work is concerned, don’t worry about people stealing your ideas. Ideas are easy, it’s the implementation that requires all the effort. In my experience, ideas fall into three categories:

- poorly thought out, not workable and not usable

- so obvious that they’ve been known for many decades

- so brilliant it will be generations before anyone understands them

The only way to distinguish among these is to find every published research paper that you can (easy with the internet) and to make everything of your own available for scrutiny.

 

 
  [ # 2 ]
Andrew Smith - Mar 11, 2013:

As daunting as the prospect might be, most of your problems would go away if you made the switch to Linux.

Ugh. I just had to cancel a programming contract I won because I couldn’t get Ubuntu working on my computer that started with Windows on it. First I struggled with a dual boot operating system but could not get wifi, so without Internet I could not download the software I was supposed to update, I needed drivers that weren’t available, I couldn’t connect to a land line, etc., so I simply was unable to do the job, so I had to write off thousands of dollars of potential income as a result. I also tried VMware for a hosted Ubuntu operating system within Windows and encountered more problems: it simply never installed after 5 hours, so I had to give up on that, too. If Linux is anywhere as problematic as Ubuntu, then Linux is not a viable option for me.

Andrew Smith - Mar 11, 2013:

Even without Linux, I know of two software packages which will run on Windows and which may benefit you a great deal in the short term.

Those sound like very good suggestions, thanks. I studied and used SVG on my own once, so that’s a good option for me.

Andrew Smith - Mar 11, 2013:

Ideas are easy, it’s the implementation that requires all the effort.

By “implementation”, do you mean hardware implementation, use of an algorithm, software simulation, any or all, or what? I’ve found what you say to be true within a commercial setting, where any new idea takes at least several weeks of programming effort to implement in software and where programmers lament a lot of new ideas for that reason, but I’m also concerned about someone maybe in an academic setting or someone maybe seeking a patent to get money and/or credit where neither money nor credit is due.

 

 

 
  [ # 3 ]

1: try open-office as a replacement for MS office. It also has a (fairly) good drawing section for creating stuff that you described (a bit like visio, but I think it is somewhat simpler, not sure), and yes Andrew Open office also runs on Windows, you don’t need linux to get Open software to run.
2:

  Andrew Smith - Mar 10, 2013:

  Ideas are easy, it’s the implementation that requires all the effort.

By “implementation”, do you mean hardware implementation, use of an algorithm, software simulation, any or all, or what? I’ve found what you say to be true within a commercial setting, where any new idea takes at least several weeks of programming effort to implement in software and where programmers lament a lot of new ideas for that reason, but I’m also concerned about someone maybe in an academic setting or someone maybe seeking a patent to get money and/or credit where neither money nor credit is due.

Andrew is right, ideas are easy to work out on paper. Programming them can be a whole different ball game, especially if you are not used to programming on a daily basis.  A few weeks for working out a proper idea in code is nothing, try a few months.

3: You can always publish in established papers. Personaly, I just put everything in a blog. If you are not called ‘google’ or ‘facebook’ (or something similar), generally nowone will care much about stealing your ideas. Only when it is working completely and you have prooved to the world (like google or facebook), how much your idea was really worth, will people start to notice, otherwise people will treat you as you have treated others on the forum here.

 

 
  [ # 4 ]

As far as claiming priority is concerned, you could always post your work on arXiv.org. (I believe it was originally developed for priority claims in the particle physics community.) The arXiv has an AI section.

The arXiv is widely read (depending on the field) but it is not a peer-reviewed publication. It’s designed for pre-prints. Most people won’t take your work terribly seriously unless you
1) are affiliated with a respected academic or corporate entity
2) publish your work in a peer-reviewed journal
3) cite the relevant literature in the field.

Apparently the first two are not currently available to you, but the third will get you far. smile

 

 
  [ # 5 ]

What I used to suggest to others, way back in the day, was what’s called a “poor man’s patent”; at least a couple of weeks before you release your work for public view, send a notorized copy of the work to yourself by registered mail in a tamper proof package. Then place the unopened package in a safe place, where it can stay safe and unmolested. In the event that you need to go to court over plagerism or other infringement concerns, you will have in your posession pre-dated proof that you had your work in your hands well prior to any dates that the opposing party can demonstrate, and since the post office handled the delivery of the package to you, that now involves a (pseudo) government agency, which gives your case even more credence.

Now please bear in mind that I haven’t had to do something like this in several decades, so this method of self-protection may not work anymore (it never had any real clout when it was first conceived, I’m sure), but even so, it’s still far better than a case of “your word against theirs”. smile

 

 
  [ # 6 ]

Tools for a windows user:
I agree with Andrew and Jan. I have used both Open Office and Inkscape on windows (along with most of the other programs that follow). They are the way to go.
What I would recommend depends on if you are doing a proof of concept/prototype or trying to create an application.
Also, HTML and Javascript can provide fast, visual examples of techniques.

PDF Creator - has the ability to print to a PDF file. As long as you can print it, it can become a PDF file.
http://www.pdfforge.org/pdfcreator

Notepad++ - my favorite editor. I use this to do AIML, Javascript, HTML
http://notepad-plus-plus.org/

Open office - This is similar to Microsoft Office.
http://www.openoffice.org/download/

Depending on what your are developing, you may be ale to prototype/demonstrate your entire application using nothing but Office.
The chat bot Harumi, is built in Open Document Format.

Incscape - This is similar to Illustrator.
http://inkscape.org/

Since you have used Python, one of the benefits of Inkscape is that you can write plug-ins in python.

Paint/Photo Editing: Similar to Photoshop
Paint.net
http://www.getpaint.net/
Has all the painting/photo editing functions you will need. You can also write C# plug-ins for it.

GIMP - The GNU Image Manipulation Program
http://www.gimp.org/
GIMP isn’t specifically targeted for windows, but there is a version available.

3D Creation:

Blender - Long learning curve, but you can use it for anything.
http://www.blender.org/
DAZ3d - Easy to use, has lots of free assets. There are some restrictions on use of output.
http://www.daz3d.com/
Unity - 3d environment
http://unity3d.com/webplayer/

Programming languages:

Python - The natural language processing course I took used this to demonstrate concepts
http://www.python.org/
The natural language tool kit (http://nltk.org/book/) has some good python examples.
In April, Coursera will be running “An Introduction to Interactive Programming in Python” course
https://www.coursera.org/course/interactivepython

Octave - similar to matlab, the neural network course I took used this to demonstrate concepts
http://www.gnu.org/software/octave/

Visual Studio Express 2012 - Free C# and web tools, you can build actual applications.
http://www.microsoft.com/visualstudio/eng/products/visual-studio-express-products

Video:
Windows Movie Maker - basic movie software
http://windows.microsoft.com/en-us/windows-live/movie-maker-get-started

Cam studio - video screen captures
http://camstudio.org/

 

 
  [ # 7 ]

’m really struggling to get some articles published, especially about my own AI architecture, especially since I don’t have access to typical software used in publications, so I could use some advice. Most of my problems stem from my low income, such as my inability to buy expensive software, inability to pay for travel expenses to present articles at conferences, inability to pay for publication fees in prestigious journals, and also from my not having an employer who will pay for such travel or fees or software. I haven’t published anything in years largely because of the effects of a bad economy, which forces me to keep changing jobs instead of doing anything of long-term benefit to anyone, namely publishing my own ideas.

Be aware, if you are trying to patent your ideas, you may NOT want to publish anything. Publishing may affect your ability to obtain a patent.  If your invention has been publicly disclosed it is not patentable.
http://www.uspto.gov/patents/process/index.jsp

Since June 8, 1995, the United States Patent and Trademark Office (USPTO) has offered inventors the option of filing a provisional application for patent which was designed to provide a lower-cost first patent filing in the United States and to give U.S. applicants parity with foreign applicants under the GATT Uruguay Round Agreements.

A provisional application for patent (provisional application) has a pendency lasting 12 months from the date the provisional application is filed.

If you think you might want to patent your invention, the preferred method of capturing your ideas may be an engineering log book. This is pretty standard in corporate organizations. Contacting a patent attorney or at minimum going through some of the resources on http://www.uspto.gov would be a good idea.

What Dave has talked about is a method of establishing “prior art”. This means you have had/used and possibly made public an idea at a time before a patent was granted. In the US patents are granted based on “first inventor”. Or, he who can demonstrate that he had the idea first. Internationally, patents are typically done on a “first filer” basis. He who submits the paperwork first gets the patent.

 

 

 
  [ # 8 ]
Jan Bogaerts - Mar 11, 2013:

Andrew is right, ideas are easy to work out on paper. Programming them can be a whole different ball game, especially if you are not used to programming on a daily basis.  A few weeks for working out a proper idea in code is nothing, try a few months.

Thanks for your feedback.

I should have been more clear. When I say I want to “publish my ideas” I mean that I plan to write professional level, proof-of-concept research papers backed up with architectural simulation results that took probably months to program, accompanied by any applicable mathematics, and *lots* of references (I love references, for some reason!). I plan to have extremely convincing demonstrations both in text and in video simulations. This is also very original material, not just say tweaks of existing formulas in learning algorithms or a new time series correlation formula. Therefore the majority of what I publish is likely to be valuable, or else I wouldn’t even bother publishing it. In contrast, I don’t plan to write some article that just lists some generic ideas like “neural networks should try incorporating Kolmogorov complexity” or “Kohonen networks should try geometrically predicting ahead for the form of the converged grid”.

 

 

 
  [ # 9 ]
Merlin - Mar 11, 2013:

Be aware, if you are trying to patent your ideas, you may NOT want to publish anything. Publishing may affect your ability to obtain a patent.  If your invention has been publicly disclosed it is not patentable.

I should have mentioned my intent there, too. I do *not* plan to get a patent. I used to work for the Patent Office, I saw what goes on behind the scenes, and I would not want anything to do with patents or the Patent Office after that. I also did a report on inventor Eli Whitney is middle school, which at an early age shattered my illusions about the benefits of patents. I also saw a chief scientist for whom I worked die and take his life’s work to the grave with him because he insisted on making money off of it via patents instead of just opting to become famous for his invention by publishing it. No, my only concern is to *prevent* anybody from patenting my ideas. If I’m not going to make any money off my ideas, I believe it’s not fair that somebody else should. I believe the key lag period for legal prevention of using a published idea for a patent is 9 months, if I remember correctly, though I’ll have to check on that.

 

 
  [ # 10 ]

I am so sincerely grateful to Dr. Amanda Windle for publishing me in her A.I. book.

 

 
  [ # 11 ]

CR:

That sounds like a great option, thanks, though I haven’t had time to look into that yet. I guess scientists aren’t supposed to be poor, or poor people aren’t supposed to be scientists, but then a lot of situations that aren’t “supposed to be” actually are.

Dave:

Yes, I heard of that self-mailed trick long ago, and it’s an inexpensive way to go even if not optimal. I think the situation should be resolvable with the Internet now, since things can be made public immediately. I’m going to make sure that any of my posted articles have exact dates for revisions on the articles themselves. It sounds like one of the tricks here is to make an article visible enough, especially via Google search, that a patent examiner can easily find it if evaluating a new patent application that attempts to copy it.

 

 
  [ # 12 ]

Have you considered teaming up with someone affiliated with a university? Most departments have resources available to pay for publication (after all, it only helps the department). And most professors have secured other sources of funding as well.

If you find someone whose work tracks yours, such a partnership could also help you get funding for your project. Write a draft for a grant proposal and see if someone is interested in putting their name on it as well, and collaborating if it’s approved. I’m not sure what your odds are, being that you aren’t in academia yourself. But it’s worth a shot—especially in computer science, where it’s more recognized that many talented people lie outside the regular channels of academia/industry.

 

 
  [ # 13 ]

Regarding A.I. articles… Any thoughts, anyone, on self publishing?

 

 
  [ # 14 ]

Nobody will take you especially seriously. If your goal is to spread your ideas to some part of the A.I. community, self-publishing won’t accomplish that. Pre-prints on arXiv can’t hurt, but that’s because it’s not an alternative to peer review.

 

 
  [ # 15 ]

With regards to patenting, also be aware that the U.S. has finally caved and become a “First to patent” system, rather than a “First to invent” system. This took effect last Saturday. In the previous system establishing your “date of inception” was important. If you could demonstrate reliably that your embodiment of a particular invention was prior or concurrent then you could successfully prosecute a claim. The new system (Which definitely favors large companies) holds that the first to file has the claim. Considering that Apple just applied for and was granted a Patent on the rectangle, and Microsoft was granted a patent for technology that we all know has existed in the prior art for years (Force sensing device for electronically transmitting touch)  the actual logic of a claim is irrelevant. Its a matter of whether or not you have the financial resources to succesfully defend against a suit brought against you by an Apple or a Microsoft. We are now definitely in the age of the “Golden Rule”

“He who has the gold, makes the rules”.

VLG

 

 1 2 > 
1 of 2
 
  login or register to react